PoEM: Predicting RL Outcomes from Existing Policies
PoEM approximates a reinforcement-learning policy for a new reward from existing post-trained models, without extra RL.
1 stories, newest first · one card per story
PoEM approximates a reinforcement-learning policy for a new reward from existing post-trained models, without extra RL.
Today 1.6K readers · 2.5K views · 779 reads · all time 189K views